基于深度学习的短期负载预测算法研究

打开文本图片集

中图分类号：H132 文献标志码：A 文章编号：1003-5168（2025)16-0029-04

DOI:10.19968/j.cnki.hnkj.1003-5168.2025.16.006

A Review of Deep Learning-Based Short-Term Load Forecasting

GE Yilin (School of Computer Science and Technology,Zhejiang Universityof Technology,Hangzhou 310014, China）

Abstract: [Purposes] The purpose of this paper is to explore the fiting degree and accuracy of short-term load forecasting with deep learning models by analyzing the principles of clasic deep learning models,and to discuss the development trends of short-term load forecasting applications based on deep learning models in the future.[Methods]This paper introduces the applications of dep learning models including Artificial Neural Network (ANN), Recurrent Neural Network (RNN), Convolutional Neural Network (CNN),and DepResidual Network (ResNet) in short-term load forecasting and analyzes the strengthsand weaknesses of those models.[Findings] The four models mentioned above allachieve accurate results in short-term load forecasting,with performances better than traditional methods.[Conclusions] Hybrid models addresssome limitations of single models and generally outperform single-model approaches in short-term load forecasting.This will be the future direction of short-term load forecasting algorithms.

Keywords: short-term load forecasting;deep learning; ANN; RNN; CNN;ResNet

0引言

准确的负载预测可以对电能的生产、传输、分类进行系统的计划与经营。（剩余7998字）